This series demonstrates how KTAP output can be used by nolibc-test to
make the test results better to read for people and machines.
Especially when running multiple invocations for different architectors
or build configurations we can make use of the kernels TAP parser to
automatically provide aggregated test reports.
The code is very hacky and incomplete and mostly meant to validate if
the output format is useful.
Start with the last patch of the series to actually see the generated
format, or run it for yourself.
Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
---
Thomas Weißschuh (7):
selftests/nolibc: statically calculate number of testsuites
selftests/nolibc: use unsigned indices for testcases
selftests/nolibc: replace repetitive test structure with macro
selftests/nolibc: count subtests
kselftest: support KTAP format
kselftest: support skipping tests with testname
selftests/nolibc: proof of concept for TAP output
tools/testing/selftests/kselftest.h | 20 +++
tools/testing/selftests/nolibc/nolibc-test.c | 197 ++++++++++--------------
tools/testing/selftests/nolibc/run-all-tests.sh | 22 +++
3 files changed, 127 insertions(+), 112 deletions(-)
---
base-commit: dfef4fc45d5713eb23d87f0863aff9c33bd4bfaf
change-id: 20230718-nolibc-ktap-tmp-4408f505408d
Best regards,
--
Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
